We investigated the relative importance of stand and landscape characteristics in the invasion of a nonnative shrub, Amur honeysuckle, in 40 woodlots in an agricultural matrix in southwest Ohio. We quantified stand characteristics that could influence invasibility, the intrinsic susceptibility of an area to invasion, including woodlot size, perimeter-to-area ratio, tree basal area, and stand age. At the landscape scale we included factors that potentially influence propagule rain (the contribution of seeds from individuals established outside the focal area), including the land cover and road density in a 1,500-m buffer around each woodlot, as well as the extent to which the perimeter was forested at two points in the past, and latitude (based on an apparent south-to-north invasion in this region). Based on stepwise regression, we determined that honeysuckle cover was determined primarily by landscape parameters, particularly the percent of the buffer comprised of cropland. Woodlots surrounded by more cropland had less honeysuckle cover, which we attribute to paucity of nearby seed sources and/or minimal movement of seed-dispersing animals. From these findings, we argue that impediments to propagule rain are more important in shaping the invasion of this exotic shrub than are characteristics of the woodlots themselves, i.e., community invasibility.

Nomenclature: Amur honeysuckle, Lonicera maackii (Rupr.) Herder.

Management Implications: Understanding the factors that make forests susceptible to invasion can inform management strategies, including identifying forest stands at greatest risk of invasion and formulating steps that can be taken to minimize invasion risk. We investigated what stand and landscape characteristics best explained the cover of the invasive shrub Lonicera maackii (Amur honeysuckle), in a landscape it recently invaded, consisting of woodlots in an agricultural matrix in southwest Ohio. We found that cover of this shrub was best explained by landscape characteristics, rather than by stand characteristics, such as age or basal area. Specifically, the percentage of the 1,500-m buffer around the woodlot that was comprised of cropland, as opposed to pasture, forest, and other land-cover types, was the best predictor of honeysuckle cover. Woodlots surrounded by more cropland had lower cover, which we think indicates more recent colonization. Thus cropland impedes honeysuckle invasion, either by providing a buffer free of seed sources (fruiting shrubs), or a land cover that is unlikely to be crossed by animals dispersing seeds from more distant sources. These findings suggest that woodlots surrounded by cropland, and perhaps by other shrub-free land covers, are at lower risk of invasion by animal-dispersed nonnative plants, and that active management of buffers around forest stands will reduce invasion risk.